【作者】 顾正华；

【导师】 唐洪武； 李云；

【作者基本信息】 河海大学 ， 水力学及河流动力学， 2004， 博士

【摘要】 本文基于智能和系统的观点，将智能科学与水科学相结合，对水流模拟的智能化问题进行了全面探索。所做研究工作的创新之处和取得的研究成果主要包括： (1)对水流研究历史作了简要回顾，论证了水流模拟智能化的必要性； (2)进一步指出水流模拟中的不确定性问题，并提出了解决途径； (3)提出水流智能模型概念，并建立了水流智能模拟的完整理论体系； (4)提出模型等价原理和水流系统智能辨识理论(包括水流系统模型智能辨识理论和水流参数智能反演理论)； (5)在基本遗传算法和加速遗传算法的基础上，提出可考虑隔代遗传的广义遗传算法； (6)对水流系统的神经网络辨识模型进行了充分研究，提出基于π定理的水流系统神经网络辨识模型、水流静态系统神经网络辨识模型、水流动态系统神经网络辨识模型，将这些模型理论实践在快速船伴流场特征量计算和透水框架式四面体空间尾流场快速仿真上，为这些问题提供了一条新的研究思路和解决途径； (7)运用水流智能模型理论对复杂河网的智能模拟进行了研究，提出具有普遍意义的河网恒定流和非恒定流智能模型，并对浦东新区河网非恒定流的实时模拟进行了研究； (8)作为工程应用，将本文创立的水流智能模型理论进一步实践在上海市浦东新区水闸智能群控信息管理系统的规划及平台建设工程中；以浦东新区水管理信息化作为工程研究背景，运用水流智能模型思想提出水闸智能调度辅助决策模型，解决了复杂河网上群闸联合调度问题； (9)对人工智能及其在水科学中的应用前景进行了探讨，并首次提出水流智能学和水流智能工程概念，为今后同类问题的研究提供了新的学术思路； (10)对本文的研究工作进行了总结，对今后的研究作了展望。更多还原

【Abstract】 On the opinions of intelligence and system, this paper probed into the problems of intelligentizing flow simulation through intelligence science combining with hydroscience. Main innovations and achievements on research include:(l)The history of flow study was investigated and necessity of intelligentizing flow simulation was demonstrated.(2) Uncertain problems in flow simulation were pointed, and approaches to solve those were presented.(3) Conceptions of Flow Intelligent Model (FIM) were defined; and system info of intelligent flow simulation was founded.(4) Principle of model equipollence and theories of flow system intelligent identification (including flow model intelligent identification and flow parameters intelligent identification) were presented.(5) On the basis of Simple Genetic Algorithms (SGA) and Accelerating Genetic Algorithms (AGA), Generalized Genetic Algorithms (GGA) was created, which can consider ancestor’s genes.(6) Intelligent identification models of flow system were fully studied based on Artificial Neural Networks (ANN). ANN identification model of flow system based on π-theorem and ANN identification models of flow static system and flow dynamic system were presented. Those model theories were applied to calculate the key variables of greyhound’s wake field and rapidly simulate penetrating frame-tetrahedron’s wake field, and provided a new idea on those problems settlement.(7) Intelligent simulation of complicated river networks was studied based on FIM theory. General intelligent models of steady flow and unsteady flow in river networks were presented, and real-time unsteady flow simulation on Pudong river networks was investigated.(8) As engineering practice, FIM theory was applied to programming and flat construction engineering of Pudong locks intelligent control information management system. On the background of Pudong water management information engineering and according to FIM’s kernels, locks intelligent control aided decision models were presented, and the problem of multi-locks united control on complicated river networks was solved.(9) Some attitudes from author were indicated about artificial intelligence and its application to hydroscience; and hydrointelligentics and Flow Intelligence Engineering (FIE) were presented for the first time and provided new academic thoughts to investigations of kindred subjects.(10) In the end, study works were summed up, and study focuses about the subject in the future were expected.更多还原